In this study, we investigated the role of Dicer protein partners in the processing of miRNA precursors. We used RNAi to silence individual Dicer protein partners and analyzed the effects of their inhibition on the endogenous and exogenous miRNA production by high-resolution northern blotting.
It has been well documented in the literature that Dicer depletion causes the accumulation of pre-miRNAs and the decrease in the amount of miRNAs (e.g. 
). It has also been shown that the level of several pre-miRNAs was not affected by Dicer depletion 
. In our experimental system, after the depletion of Dicer by RNAi, we observed reduced levels of mature miRNAs and elevated levels of their precursors. Therefore, our observation is in agreement with the majority of previous reports (Table S1
). We also detected the decrease in miRNA amount upon AGO2, TRBP, and PACT depletion. Similar results were reported in the literature and explained by the reduced efficiency of dicing caused by a deficiency of Dicer protein partners 
. However, upon AGO2, TRBP and PACT depletion we also observed the decrease in the cellular levels of miRNA precursors, which has not been reported so far. As the deficiency of protein partners of Dicer influences the level of Dicer itself (, and ), we hypothesize that the pre-miRNA decrease may result from the decreased cellular levels of the functional RLC and a rapid degradation of pre-miRNAs that did not enter the processing complex. The fact that the reduction of TRBP protein level reduces also the level of Dicer protein is in good accordance with the results of others (Table S1
), suggesting that the lack of TRBP decreases Dicer stability and/or affects the assembly of the Dicer complex 
. Similarly, upon PACT depletion the fall in the Dicer level was observed 
. It is not surprising that the knockdown of TRBP and PACT causes similar effects, as these proteins can interact with Dicer in a similar manner and can act redundantly to some extent 
. As far as AGO2 depletion is concerned, the observed decrease in pre-miRNAs' level is a novel finding of this study. Earlier, it has been shown that AGO2, which also has a PAZ domain, can bind and cleave pre-miRNAs 
. It was also proposed that AGO2 may facilitate the access of pre-miRNAs to Dicer 
. Therefore, the AGO2 binding may contribute to the protection of pre-miRNA from unspecific degradation.
We have performed the siRNA-mediated knockdown of Dicer protein partners and not their knockout, therefore some portion of these proteins remained in the cells. These proteins could form a smaller amount of the fully functional Dicer complex. Therefore, only a portion of pre-miRNAs could be processed and the rest underwent degradation. As a result, we observed smaller amounts of pre-miRNAs and miRNAs.
Our experiments revealed that the silencing of Dicer protein partners had mostly a minor effect on the specificity of Dicer cleavages. The distribution of the length variants of miRNAs differed slightly between the analyzed samples (siRNA targeting AGO2, TRBP and PACT) and the control (siRNA LUC). The changes in the Dicer cleavage pattern could result from the activity of the RNase fraction having changed cleavage specificity when deprived of either protein partner.
The transfection of synthetic pre-miRNAs combined with the northern blotting analysis may be considered another experimental approach to study the Dicer step of miRNA biogenesis. This system, however, needs improvement to make it applicable to a larger number of pre-miRNAs. The synthetic miRNA precursors, used in these experiments, differ in their sequence and structure and may not be equally good substrates for binding and cleavage by the endogenous Dicer complex. It is unlikely that such precursors are fully captured by Dicer. Their substantial portion may be trapped in endosomes. The fraction that escapes endosomes may enter the miRNA biogenesis pathway or may become a target for cellular exo- and endoribonucleases. We do not know what the contribution is of such unspecific processes to the miRNA patterns observed in . We propose, however, that this contribution is not very large because most of the observed northern blot signals belonged to undegraded precursors and miRNA sized products. Recent research shows that siRNAs transfected to cells are primarily intercepted by Dicer, which recognizes 2nt 3′-overhangs, but the efficiency of capture of different siRNAs is highly variable and dependent on the nature of the 3′ overhang sequence 
. It is also likely that synthetic miRNA precursors transfected to cells may be primarily captured by Dicer that is present in the RLC complex, but the affinity of the Dicer complex to different pre-miRNAs may vary greatly.
The mechanism of Dicer cleavage implies that on the way to the final miRNA-miRNA* duplex resulting from cleavages in both precursor arms by two RNase III domains, the intermediate product may be generated resulting from a more rapid cleavage in one precursor arm by one of the RNase III domains of Dicer. The single-nick intermediate is observed in various amounts for various substrates when synthetic precursors are cleaved with recombinant Dicer 
. In contrast, in HeLa cells transfected with synthetic pre-miRNAs or miRNA-encoding plasmids, the intermediate product was not observed, suggesting the potential role of protein partners in the synchronization of cleavages generated by individual RNase III domains of Dicer. We considered two alternatives to explain the easy detection of the single-nick intermediates in reactions with recombinant Dicer and the inability to detect this cleavage intermediate in cellular systems. One possibility is that the endogenous Dicer and its protein partners may release the single-nicked precursor to the cytosol, where it undergoes rapid degradation by ribonucleases and escapes detection. The possibility which we consider more likely is that the cleavages generated by two RNase III domains of Dicer are much better synchronized when a precursor undergoes cleavage within RLC than when it is cleaved by recombinant Dicer alone.